Why do rational people vote in large elections with costs to vote Serge Blondel GRANEM Univ. Angers

1
Why do rational people vote in large elections
with costs to vote? Serge Blondel (GRANEM Univ.
Angers INH, CES)Louis Lévy-Garboua (CES
University Paris 1)
IAREP-SABE 08 Rome
2
Voting dec Non-EU Examples
American presidential election Bush-Gore each
vote in Florida has counted. This issue is very
rare, but this elcetion has recalled that it is
possible.
This result has not changed the behavior of
people for who voting is a citizen obligation
because this is the foundation of a democracy.
We will study here the vote as a decision, as an
alternative to abstention.
1
3

• Voting dec
• Hyp.
• Decision
• Behaviour
• Non-EU
• Examples

Hypotheses NOT considered - Taste for voting -
Game theory solution ? Individual decision
framework
Hypotheses considered here - Preference for one
candidate - Cost of voting - Probability of
being decisive e e1 ½ e2 ? e1 equality if
no vote? e2 equality if vote
2
4

• Voting dec
• Hyp.
• Decision
• Behaviour
• Non-EU
• Examples

Owen Grofman (84)
3
5

• Voting dec
• Hyp.
• Decision
• Behaviour
• Non-EU
• Examples

EU V ? A ? (1-e).C lt e.(B-C) ? e.B gt
C Impossible since e0
Paradox of not voting (PNV) Downs 57
4
6

• Voting dec
• Hyp.
• Decision
• Behaviour
• Non-EU
• Examples

Observed behaviours - People do vote - The rate
of participation increases with the stake of the
election - The rate of participation increases
with the uncertainty of the election
5
7

• Voting dec
• Non-EU
• Regret th.
• Prospect th.
• CC th.
• Examples

Paradoxes and anomalies of EU - Gambling and
insurance (Friedman Savage, 1948) - Allais
paradox (Allais, 1953) - Preference reversal
phenomenon (Lichtenstein Slovic, 1971) -
Reflection effect (Kahneman Tversky, 1979)...
gt non-EU theories (Starmer 2000) related to the
previous results, not the PNV.
6
8

• Voting dec
• Non-EU
• Regret th.
• Prospect th.
• CC th.
• Examples

N 30 millions and q50 gt e 0.03 We pose -
B 10 - C 1
Two hypotheses ignored by EU H1 utility
depends of the outcome and the outcome in the
alternative decision H2 e should be
overestimated
7
9

• Voting dec
• Non-EU
• Regret th.
• Prospect th.
• CC th.
• Examples

Ferejohn Fiorina (74) Loomes Sudgen
(82) Regret theory EU regret V ? A ? 4.0015
0.9997R(-1) 0.0003 R(9) gt 4.9985
0.9997R(1) 0.0003 R(-9) H1, not H2 impossible
8
10

• Voting dec
• Non-EU
• Regret th.
• Prospect th.
• CC th.
• Examples

w(p)
1
0
p
1
Kahneman Tversky (79) Quiggin (82) Prospect
theory V ? A ? w(0.50015).9 - w(0.49985) gt
w(0.49985).10 H2, not H1 impossible
9
11

• Voting dec
• Non-EU
• Regret th.
• Prospect th.
• CC th.
• Examples

Cognitive consistency theory Lévy-Garboua 99
decision in two steps 1/EU maximisation 2/ Focus
on the objection to EU
Weighted average of EU objection V ? A ?
µ.4.0015 (1-µ).9 gt µ.4.9985 (1-µ).0 ,
0ltµ1 ? µlt0.9003 H1 and H2 possible
10
12
Voting dec Non-EU Regret th. Prospect th.
CC th. Examples
General problem V ? A ? µ.(eB-C) (1-µ).(B-C)
gt 0 Prob (V ? A) - Increases with B -
Decreases with C - Increases with e - Decreases
with µ
An objection to this theory e has a
negligeable influence on the criterion. ? The
impact of e is marginal in relation with B, C and
µ. ? The weight of the objection is independent
from e
11
13

• Voting dec
• Non-EU
• Regret th.
• Prospect th.
• CC th.
• Examples

V ? A ? µ.(eB-eC-(1-e)C) (1-µ).(B-C) gt 0 ?
(1-µ(1-e)).(B-C) µ.(1-e)(-C) gt 0
w(1-e) w(e)
The cognitive consistency theory explains the
PNV and also the main paradoxes and anomalies of
EU with only one extra parameter.
12
14

• Voting dec
• Non-EU
• Examples

13
15

• Voting dec
• Non-EU
• Examples

THANK YOU for your attention ! Questions?
14